半参数方法对上证指数波动率的预测分析

根据金融时间序列一般都存在条件异方差性,本文研究了两种半参数时间序列模型的估计方法,通过实际算例对参数模型与半参数模型,以及不同的半参数模型在金融时间序列的拟合与波动率预测方面的表现作了比较.结果表明,半参数模型要优于参数模型,半参数可加模型要优于多项式样条估计模型.     

Oxide transformation and break-up of liquid metal in boiling solutions

The non-toxic gallium-based liquid metals(LMs) are promising for various newly emerging areas due to having excellent fluidity, high electrical and heat conductivity. However, the oxide layers often affect their applications. In this paper, we disclosed a straightforward boiling strategy to efficiently remove the oxide of LM in deionized water. Meanwhile such method was also demonstrated to be capable of quickly breaking up LM into droplets(mean 0.51 mm, SD 0.20 mm) in Na Cl solution,which has important value for manufacture purpose. According to the force analysis, surface tension of LM dominates the interaction between solution bubbles and LM under the same heating conditions. Then we characterized the surface morphology and compositions of LM droplets boiled(LMDB) using scanning electron microscope(SEM), X-ray diffraction(XRD) and Xray photoelectron spectroscopy(XPS) and tested the dynamic force between a rod and LMDB. It was shown that after being boiled in water, most of Ga_2O_3 was transformed to scattered α-Ga OOH nanorods which has less restriction to the deformation of LM and large resistance force(0.8 mN) was tested; otherwise, after boiled in Na Cl solution, more Ga2 O3 with contact patch structure reduces the surface resistance of LM(0.2 m N) and makes it to be broken up by bubbles. The different surface products are determined by the inhibition of hydrolysis of Ga~(3+)by Na Cl solution. This study paved a simple way for oxide removal in neutral solution and mass production of LMD. It also suggests an efficient approach for exploring mechanisms of interactions between LM and bubbles, which is difficult to study otherwise in sonication.